BACKGROUND: Rising UCL injury rates at both amateur and professional levels have been linked to fatigue in baseball pitchers. Repeated pitching has been associated with changes to kinematics, kinetics, and perceived fatigue, but no statistical model exists which incorporates all the most common aspects of fatigue into one framework. Confirmatory factor analysis (CFA) can be used to investigate possible fatigue frameworks and their plausibility in explaining the multivariate nature of observed changes occurring with repeated pitching. AIM: To investigate how multiple fatigue manifestations could be associated with a shared factor in baseball pitching. METHOD: Two CFA models were proposed; one a priori model based solely on previous findings and literature linking commonly found variables which change with repeated pitching, and one a posteriori model with added correlation factors between maximum external shoulder rotation (MER) with perceived fatigue and MER with triceps EMG activity. RESULTS: Model fitness test performed on the first a priori model proved plausible, with it passing some of the tests but failing others. The a posteriori model showed to be an excellent model for explaining the covariance of the data, passing all model fitness tests. CONCLUSIONS: Confirmatory factor analysis can serve to provide a plausible framework for explaining the covariance measured in kinematic, kinetic, and other fatigue related changes to baseball pitching data. Both models would suggest that the shared latent variable represents an underlying aspect of physiological fatigue. Changes to MER were determined to not be directly caused by fatigue. A proper understanding of different fatigue manifestations can potentially reduce the amount of fatigue related UCL injuries plaguing baseball pitchers by providing a more accurate proxy for measuring physiological fatigue.